NECC: Enabling Non-Expert Customized Congestion Control with LLMs

Mingrui Zhang, Hamid Bagheri, and Lisong Xu present NECC, a framework that enables non-expert users to model, implement, and deploy customized congestion control algorithms using Large Language Models and the Berkeley Packet Filter (BPF) interface.

• Main announcement: The authors introduce NECC, an exploratory non-expert customized CCA framework that leverages Large Language Models and the Berkeley Packet Filter (BPF) interface to let non-expert users model, implement, and deploy customized congestion control algorithms; the work is published on arXiv:2601.22461 and is the accepted manuscript (AAM) for IEEE ICC 2025 (Related DOI: https://doi.org/10.1109/ICC52391.2025.11160790).
• Background/details: The paper reports promising evaluation results using real-world CCAs, provides PDF/HTML/TeX source on arXiv, and lists submission metadata [v1 submitted 30 Jan 2026]; the entry references DataCite (arXiv-issued DOI pending registration).